Asymmetric Volatility Spillover between Stock Market and Foreign Exchange Market: Instances from Indian Market from Pre-,during and Post- Subprime Crisis Periods

Abstract

Modern businesses are so inter-twined that a cause in one market affects other markets throughout the Globe. The 2008 subprime crisis is one of such evidences of inter-linkage of global markets. Such type of event motivates many studies to analyse the transmission of volatility from one market to another market. The study aims to analyse the volatility spillover effect between CNX Nifty and exchange rates covering for three different currencies, that is, USD, GBP and yen. GARCH (1,1) and EGARCH (1,1) have been used to identify the spillover effect and asymmetries or leverage effect in the volatility transmission through the estimation of different parameters. The overall findings show that there is spillover between the foreign exchange and the stock market. Among the three exchange rates, the USDR is strongly co-related with the Indian stock market as compared to other rates. Our study will significantly contribute to the existing literature in this context. The findings of the study have greater implications especially for hedgers, arbitrators and other participants in this market. As such type of information regarding transmission of volatility can help them to diversify their overseas risk through an optimal portfolio selection.

Keywords

Volatility clustering and pooling volatility spillover leverage effect CNX Nifty exchange rates

Introduction

The corporate world has witnessed a number of financial debacles in past and came out, somewhat, little scratched from their wounds. But the recent subprime crisis of 2008 is one of the scariest and deadliest crises till date as it has vanished out many of the big financial players, who were ‘too big to fall’, in its financial blood-bath. This financial crisis spread like Tsunami from one continent to another and from one nation to the other. This crisis has also questioned the linkage and asymmetric cross-market volatility spillover between the markets. Many of the studies are since then trying to find out the spillover effect during the crisis period. However, it is believed that the subprime crisis did not affect the Indian market directly to a great extent, yet our economy was indirectly affected because of the flight of capital. The major steps taken by our policy makers for financial stability helped the Indian economy in quick recovery (Ghosh, 2014). In another context, liberalization of foreign exchange policies and adoption of market-driven price determination system has increased the cross-border capital flows between the nations. This greater capital flows have raised the curiosity about the transmission of volatility effect and inter-relationship between equity and foreign exchange market. For example, if the investors of the US will invest in the Indian stock market, then the demand for our home currency increases and thus the exchange rate strengthens and vice versa. Thus, the volatility in exchange rate drives the investment strategy and leads the trading pattern of the potential investor.

The present study analyses the volatility spillover between Indian equity and foreign exchange market. We have taken CNX Nifty return and three exchange rates, namely USD, GBP and yen. We use GARCH (1,1) and EGARCH (1,1) to study the asymmetric volatility spillover across the market. Our study period has been divided into three sub-periods such as pre, during and post-crisis periods. The remaining sections of the present study can be divided as follows. The second section deals with the literature review to the concerned subject matter. The next section describes the objective of the study. The fourth section explains the data and methodology. The subsequent section presents empirical results of the present article in which descriptive statistics and volatility spillover between stock and exchange rates have been covered in general, and spillover effect has been observed between stock and USDR, stock and GBPR, and stock and YENR, in particular. The sixth section contains conclusion to the study and the final section explains managerial implication.

Literature Review

Plentiful number of studies have already been made in the context of relationship between the stock market and the foreign exchange market. For instance, Francis, Hunter, and Hasan (2002) analysed volatility between stock and currency market. The findings of the study showed that equity market drives the volatility of currency market and also the past volatility has impact on the current volatility. In this context, another study of Bonga-Bonga and Hoveni (2013) has found that there is one-way volatility transmission from equity to forex market. By studying exchange rates of six countries, that is, the US, the UK, Japan, France, Canada and Germany, Kanas (2000) found evidence of volatility spillover from stock market to exchange rate in all the countries except in Germany. Similar kind of studies have been conducted not only in the foreign market, but also in the context of Indian emerging market. For example, Apte (2001) has studied the transmission of volatility between the Indian stock market and INR–USD exchange rates, and using EGARCH, the article found volatility spillover from stock to exchange rate. In addition, the study argues that as per the proposition of CAPM model, the volatility in forex market can be considered as a systematic risk which can be hedged by diversification in foreign exchange market. Therefore, the study of relationship is much more important. In another study, the asymmetric volatility has been found out between stock and forex market (Badrinath & Apte, 2005, pp. 1–26). The study conducted by Saha and Chakrabarti (2011) analysed the asymmetric spillover between stock and foreign exchange market. This study showed volatility spillover between stock and foreign exchange market in India. Another study has been conducted by Sahoo (2012), in which the researcher found evidence of volatility transmission from Indian rupee to other foreign currencies like dollar, pound sterling, yen, etc. Similarly, Gupta and Kamila (2015) studied the link between the implied volatility indices. The study has taken two groups of sample like for developed market, the US, the UK and Japan, and for emerging market, the four BRIC countries. The article concluded that US IVI has significant impact on other IVIs. Zabiulla (2015) analysed the volatility pattern and leverage effect in case of exchange rates. The study has taken the sample from 3 January 2000 to 31 December 2012 and concluded that there was no leverage effect; however, volatility clustering was there. This type of study not only involves different markets, but also the researchers have analysed this in different time phases. If the conditional correlation varies over time, then there is a possibility of portfolio diversification and adopting risk minimizing hedging strategies (Kumar, 2014). In the context of subprime crisis, Panda and Deo (2014) have studied the transmission of volatility between equity and foreign exchange market. The study analysed the spillover during pre and post subprime crisis and found that the evidence of bi-direction volatility spillover is more volatile during post-crisis period. In another study, contagion effect has been analysed by Chittedi (2015). The article used DCC GARCH by taking a sample from January 2002 to December 2011. Further, the study divides the pre- and the post-crisis periods. The conclusion shows that there is contagion between the US and the Indian market. Also the crisis period as compared to the pre-crisis period, shows a higher correlation. Srivastava, Bhatia, and Gupta (2015) analysed the interaction between the US and the other emerging Asian stock markets before and after the global financial crisis. By taking the sample from January 1992 to April 2014, the study concludes that there is long-term integration between Indian and the Global market; however, the study showed evidence against short-term integration.

Among the numerous studies, it has been observed that though many studies have been already undertaken in this context, very few have observed the volatility spillover during the subprime crisis. Also the conflict in results always provides an interesting scope for further research. The present study differs from the past studies in many contexts like division of sample period into three groups, and also the present study selects three major exchange rates for better understanding of contagion between stock and foreign exchange market.

Objective of the Study

Our study intends to analyse the volatility clustering and persistence in the stock and foreign exchange market. Also we have analysed the pattern of volatility transmission and asymmetries between stock prices and exchange rates. The study specifically analyses the volatility spillover in pre, during and post subprime crisis periods. So that, the immediate impact of global financial crisis on volatility transmission can be observed.

Data and Methodology

For the purpose of analysis, daily returns of CNX Nifty and three exchange rates, namely USD, GBP and yen have been taken from the Bloomberg data base. We have taken daily returns, as daily data can exhibit better volatility clustering and spillover. The period of the study spans from 1 April 2007 to 31 March 2015. The common sample has been selected by deletion of the whole period in case of missing of any one data. Further, the period of study has been divided into three sub-periods such as pre, during and post-crisis periods. As the starting of the crisis period can be traced to the declaration of bankruptcy of Lehman brothers on 15 September 2008, the period before this is considered as pre-crisis period. Economist believes that the crisis continued up to the end of October of the year 2009, so crisis period is up to 31 October 2009 and after that the post-crisis period (Ghosh, 2014). Financial crisis is a situation that creates panic in the economy and leads to bubble burst and stock market crash. Many investors lose their money and economic slowdown occurs because of recession in the market. Major financial institutions and stock markets also collapse and it causes economic meltdown. The subprime financial crisis mainly occurred due to housing bubbles and banking emergencies which started from the US markets, spreading all over the world. The major concern regarding such financial crisis is its magnitude of innovation in financial market over the period of time, and how long the impact persists in the market. In order to assess such contagion, the whole sample period has been classified as follows:

Pre-crisis period—from 1 April 2007 to 14 September 2008.

During crisis period—from 15 September 2008 to 31 October 2009.

Post-crisis period—from 1 November 2009 to 31 March 2015.

The natural logs of the entire variable have been taken to make the series log linear. The return of different exchange rates like USD changes (USDR), yen changes (YENR) and GBP changes (GBPR) and stock market (Stock) has been calculated as follows, where P_t is price at t period, P_t_–1 is price at t–1 period and Ln is natural log:

R_{i t} = L n (\frac{P_{t}}{P_{t - ​ 1}})

The methodologies of studies are as follows:

Test of Stationarity: Before applying any test, the data should be stationary. So first of all, stationarity has been checked on the basis of Augmented Dickey–Fuller (ADF) and Phillips–Perron (PP). Stationarity means the series does not have unit root problem, so there will not be any problem in statistical inferences.

Volatility Clustering and Spillover: We have employed the ARCH family models: ARCH of Engel (1982); GARCH of Bollerslev (1986) and EGARCH of Nelson (1991). If the GARCH is significant, it means previous return volatility can influence the current return volatility. Apart from past innovation, the dependency of conditional variance upon its own previous lag can be assessed in GARCH. The primary motivation behind use of ARCH family model is, it captures volatility clustering or pooling, as volatility clustering describes the positive correlation between the current level volatility with immediate past level. Assume that the stock return (ΔSP_t) and exchange rate changes (ΔEX_t) are represented by AR(1) process, then in both GARCH(1,1) and EGARCH(1,1) equation will be as follows:

Δ S P_{t} = α + β Δ S P_{t - 1} + ς_{t}, ς_{t} ~ N (0, σ_{t, (Δ S P)}^{2})

(1)

Δ E X_{t} = α + β Δ E X_{t - 1} + ε_{t}, ε_{t} ~ N (0, σ_{t, (Δ E X)}^{2})

(2)

In Equations 1 and 2, α is the intercept, SP_t–1, EX_t–1 are the returns of stock price and exchange rate, respectively, in t–1 period. Similarly, ς_t and ε_t are the white noise error terms of stock price and exchange rate, respectively.

Traditionally, standard deviation is used to assess unconditional variance, that is, crude measure of variance. While GARCH can be used to explain the conditional variance, that is, true measure of variance. We have employed the same equation as given by Panda and Deo (2014). The spillover equation in GARCH (1,1) model can be explained as follows:

h_{t (S t o c k I n d i c e s)} = α_{1} + β_{1} \in_{t - 1}^{2} + β_{2} h_{t - 1} + θ (s q u a r e d r e s i d u a l s o f e x c h a n g e r a t e s)

(3)

h_{t (E x c h a n g e R a t e)} = α_{1} + β_{1} \in_{t - 1}^{2} + β_{2} h_{t - 1} + θ (s q u a r e d r e s i d u a l s o f s t o c k I n d i c e s)

(4)

In both Equations 3 and 4, the parameters α₁, β₁, β₂ ≥ 0. Here h_t indicates the conditional variance. Here the lag of squared residuals from the mean equation $ε_{t - 1}^{2}$ and previous variance h_t _–1 explains the h_t , while θ is spillover. Further to test the asymmetric or leverage effect, that is, the impact of good or bad news, the EGARCH model has been used. In EGARCH model, the spillover equation will be as follows:

\begin{matrix} \ln h_{t (S t o c k I n d i c e s)} = α_{1} + β_{1} \ln h_{t - 1} + β_{2} | \frac{ε_{t - 1}}{\sqrt{h_{t - 1}}} | \\ + γ | \frac{ε_{t - 1}}{\sqrt{h_{t - 1}}} | + θ (r e s i d u a l s o f e x c h a n g e r a t e s) \end{matrix}

(5)

\begin{matrix} \ln h_{t (E x c h a n g e R a t e)} = α_{1} + β_{1} \ln h_{t - 1} + β_{2} | \frac{ε_{t - 1}}{\sqrt{h_{t - 1}}} | \\ + γ | \frac{ε_{t - 1}}{\sqrt{h_{t - 1}}} | + θ (r e s i d u a l s o f S t o c k I n d i c e s) \end{matrix}

(6)

In Equations 5 and 6, ln h_t _{(
Stock Indices
)} and ln h_t _{(
Exchange Rate
)} are natural log of conditional variance of stock and exchange rate, respectively. b ₂ denotes the impact of change in news on volatility and c shows the leverage effect. i is the volatility spillover between both the markets. c represents the asymmetric effect in the model. Here the coefficient may be negative or positive; the positive shows the shock innovation of good news and negative shows the shock innovation of bad news.

Analysis

Descriptive Statistics

Table 1 shows the descriptive statistics of stock return for all the sub-periods. In all the sub-periods, the average return is positive. And the standard deviation shows the volatility, it is more during the crisis period. Similarly, in both before and after crisis period, the series is negatively skewed; while during the crisis period it is positively skewed. Also the peakedness of the series shows leptokurtic of the series. The Jarque–Bera is significant at 1 per cent level; so hypothesis of normality of distribution has been rejected.

Table 1.

Descriptive Statistics of Stock Return

Period	Mean	Median	Max.	Min.	SD	Sk	Kurto	J–B
Pre-crisis	0.000418	0.001005	0.067205	–0.091046	0.019945	–0.263122	4.965524	62.62076***
During the crisis	0.000737	0.001179	0.163343	–0.13014	0.026938	0.293854	8.777787	403.3333***
Post-crisis	0.000393	0.000458	0.037380	–0.04169	0.010776	–0.01375	3.905925	44.05056***

Source: Authors.

Notes: SD indicates standard deviation, Sk indicates skewness, Kurto indicates kurtosis and J–B indicates Jarque–Bera. *** denotes significance at 1% level.

Table 2 shows the descriptive statistics of USDR. In all the three sub periods, the average return is positive. Standard deviation shows the volatility in different periods; it can be observed that volatility is high during the crisis period. Similarly, in all the periods, the series is positively skewed. And the peakedness of the series shows leptokurtic of the series. The Jarque–Bera is significant at 1 per cent level; so the hypothesis is rejected and data are abnormal.

Table 2.

Descriptive Statistics of USDR

Period	Mean	Median	Max.	Min.	SD	Sk	Kurto	J–B
Pre-crisis	0.000139	3.27	0.014382	–0.01388	0.003798	0.073781	4.620092	40.02792***
During the crisis	0.000439	–5.25	0.028605	–0.03029	0.006788	0.083484	5.104655	53.30367***
Post-crisis	0.000233	0.000115	0.039038	–0.03294	0.005542	0.234171	8.194567	1,458.754***

Source: Authors.

Notes: SD denotes standard deviation, Sk denotes skewness, Kurto denotes Kurtosis and J–B indicates Jarque–Bera. *** denotes significance at 1% level.

The descriptive statistics of GBPR has been presented in Table 3. In both pre and during crisis periods, the average return is negative, while it is positive in post-crisis period. Similar to USDR, the GBPR is also more volatile during the crisis period. Skewness is positive in both before and after crisis periods. Also the peakedness of the series shows leptokurtic of the series. The Jarque–Bera is significant at 1 per cent level. Therefore, the hypothesis of normality of distribution has been rejected.

Table 3.

Descriptive Statistics of GBPR

Period	Mean	Median	Max.	Min.	SD	Sk	Kurto	J–B
Pre-crisis	–0.000159	–0.000399	0.022306	–0.015454	0.005422	0.287119	4.183797	26.18328***
During the crisis	–0.00024	0.000615	0.035969	–0.03946	0.011310	–0.70052	4.489806	50.01479***
Post-crisis	0.000146	0.000242	0.035637	–0.02883	0.005967	0.018353	5.019104	218.6897***

Source: Authors.

Notes: SD indicates standard deviation, Sk denotes skewness, Kurto indicates Kurtosis and J–B indicates Jarque–Bera. *** denotes significance at 1% level.

Table 4 depicts the descriptive statistics of YENR. In both during the pre-crisis period and during the crisis period, the average return is positive, while negative during post-crisis period. YENR also shows similar volatility pattern as in case of USDR and GBPR. During before and after crisis period, the series is positively skewed, while during the crisis period it is negatively skewed. Also the peakedness of the series shows leptokurtic of the series. The Jarque–Beta is significant at 1 per cent level, so normality of distribution has been rejected.

Table 4.

Descriptive Statistics of YENR

Period	Mean	Median	Max.	Min.	SD	Sk	Kurto	J–B
Pre-crisis	0.000394	–0.000597	0.052861	–0.026026	0.009359	0.957219	7.385160	346.2823***
During the crisis	0.000730	0.001187	0.057050	–0.04699	0.013491	–0.0361	4.325465	21.07144***
Post-crisis	–0.000174	–0.00017	0.038424	–0.03651	0.008212	0.154907	5.356088	302.8275***

Source: Authors.

Notes: SD indicates standard deviation, Sk indicates skewness, Kurto indicates kurtosis and J–B indicates Jarque–Bera. *** denotes significance at 1% level.

Unit Root Test: Before proceeding for further test, the stationarity of all the variables contained in present study has been checked across all the period i.e., Pre-Crisis, during crisis, and Post-crisis. Table 5 shows the result of unit root test on the basis of ADF and PP. In case all the variables and during all the periods the variables are stationary at level, as null hypothesis has been rejected at 1 per cent level.

Table 5.

Unit Root Test

Series	Pre-crisis		During the Crisis		Post-crisis
Series	ADF	PP	ADF	PP	ADF	PP
SXR	–18.55797***	–18.55734***	–15.59079***	–15.54981***	–33.17559***	–33.10161***
USDR	–18.55797***	–17.29256***	–15.69411***	–15.75587***	–27.15702***	–34.70936***
GBPR	–18.40674***	–18.45096***	–13.29780***	–17.04465***	–26.19150***	–33.23278***
YENR	–19.95901***	–19.96085***	–15.96628***	–16.35061***	–34.76863***	–34.76181***

Source: Authors.

Note: *** indicates significance at 1% level.

Volatility Spillover between Stock and Exchange Rates

The results presented in the following tables show all the coefficients of GARCH and EGARCH models. The results are shown in three panels: the first panel ws the results of volatility spillover between stock and USDR, the second panel shows spillover between stock and GBPR and the third panel shows spillover between stock and YENR. In all the tables, α₀ and α₁ are the parameters of mean equation, and parameters of variance equations are

b _{0 =} Constant in variance equation

b _{1 =} Lagged squared residuals (ARCH term)

b _{2 =} Lagged conditional variance (GARCH term)

γ = Leverage effect

θ = Spillover effect

Volatility Spillover between Stock and USDR

Table 6 shows the results during the pre-crisis period, Table 7 shows results during the crisis period and Table 8 shows results during the post-crisis period. In Table 6, in both GARCH (1,1) and EGARCH (1,1) models, β₁ and β₂ are highly statistically significant. It indicates presence of both volatility clustering and persistence. In both equations, the parameter θ is also statistically significant, so there is bi-directional volatility spillover between stock and USDR. Also the spillover is asymmetric as the coefficient is statistically significant. While during the crisis period (Table 7), β₁ is significant in both equations and β₂ is significant only in GARCH (1,1) model between USDR and stock, in other case it is insignificant. In GARCH model, there is spillover from stock to USDR unidirectionally, while it contradicts in EGARCH (1,1), that is, bi-directional. However, there is no leverage effect. In the post-crisis period (Table 8), both ARCH and GARCH terms are statistically highly significant. But there is only unidirectional spillover from USDR to stock in both equations and also there is leverage effect. The γ is positive in case of stock to USDR and negative in case of USDR to stock. In all the three periods, ARCH-LM diagnosis test shows there is no ARCH effect as it is not significant.

Table 6.

Volatility Spillovers between Stock and USDR during the Pre-crisis Period

Coefficients	GARCH (1,1)		EGARCH (1,1)
Coefficients	Stock → USDR	USDR → Stock	Stock → USDR	USDR → Stock
α ₀	–5.13	0.001336	4.65	0.000828
α ₁	–0.074***	–1.87344***	–0.06936***	–1.877811***
β ₀	5.31***	2.14***	–1.2702***	–1.016256***
β ₁	0.161437***	0.164708***	0.341667***	0.230421***
β ₂	0.801312***	0.77579***	0.912236***	0.898338***
γ			0.063506*	–0.180127***
θ	–2.83***	0.003973**	–4.79842***	21.49770**
ARCH LM	0.033070	0.004431	0.027618	–0.044982

Source: Authors.

Notes: ***, ** and * indicate significance at 1%, 5% and 10% levels, respectively.

Table 7.

Volatility Spillovers between Stock and USDR during the Crisis Period

Coefficients	GARCH (1,1)		EGARCH (1,1)
Coefficients	Stock → USDR	USDR → Stock	Stock → USDR	USDR → Stock
α ₀	0.000153	0.001072	8.45	0.00038
α ₁	–0.15304***	–2.49613***	–0.16108***	–2.36422***
β ₀	1.86***	–4.40	–5.97141***	0.015962***
β ₁	0.301591***	–0.01661***	0.331643***	–0.05253***
β ₂	0.093953	1.009944***	0.454788***	0.998065***
Γ			–0.02843	0.013919
Θ	–0.00012***	–0.00121	–9.25854***	3.823757***
ARCH LM	–0.028105	–0.014378	0.022446	–0.028336

Source: Authors.

Notes: ***, ** and * indicate significance at 1%, 5% and 10% levels, respectively.

Table 8.

Volatility Spillovers between Stock and USDR during the Post-crisis Period

Coefficients	GARCH (1,1)		EGARCH (1,1)
Coefficients	Stock → USDR	USDR → Stock	Stock → USDR	USDR → Stock
α ₀	0.000165	0.000682***	0.000259**	0.000517**
α ₁	–0.20001***	–0.91751***	–0.20239***	–0.89953***
β ₀	4.66***	1.98***	–0.29743***	–0.36727***
β ₁	0.076555***	0.039449***	0.120379***	0.079055***
β ₂	0.903149***	0.93706***	0.981185***	0.96757***
Γ			0.058689***	–0.0627***
Θ	–1.50	0.000624***	–0.40044	6.599427***
ARCH LM	–0.005188	–0.025110	0.007908	–0.020185

Source: Authors.

Notes: ***, ** and * indicate significance at 1%, 5% and 10% levels, respectively.

Volatility Spillover between Stock and GBPR

In Table 9 for spillover between stock and GBPR during the pre-crisis period, the GARCH term is significant but the result is surprising that there is no spillover effect between stock and GBPR in both GARCH (1,1) and EGARCH (1,1) equations. While in Table 10, the results show that there is unidirectional spillover from stock to GBPR only in EGARCH (1,1) equation during the crisis period. And during the post-crisis period (Table 11) both the ARCH and GARCH terms are significant, except GARCH term in EGARCH model between stock and GBPR. So there is both volatility clustering and persistence in the series. Spillover co-efficient shows that there is unidirectional spillover from stock to GBP in both equations. This contradicts both pre and during the crisis period results. Finally, the ARCH LM test shows that there is no ARCH effect in the error terms. The results clearly indicate that financial crisis has some impact on volatility spillover, as the pre-crisis period shows no spillover, while during and after the crisis period signifies unidirectional spillover.

Table 9.

Volatility Spillovers between Stock and GBPR during the Pre-crisis Period

Coefficients	GARCH (1,1)		EGARCH (1,1)
Coefficients	Stock → GBPR	GBPR → Stock	Stock → GBPR	GBPR → Stock
α ₀	–0.00017	0.001594*	–0.00017	0.000876
α ₁	–0.05179***	–0.79696***	–0.0464***	–0.73425***
β ₀	1.72	1.47***	–0.00561	–1.26605***
β ₁	0.032877*	0.155725***	–0.00806	0.169213***
β ₂	0.964061***	0.816271***	0.998752***	0.869435***
Γ			0.051445***	–0.23815***
Θ	–1.58	–0.00133	–0.04077	–5.83298
ARCH LM	–0.055523	–0.003623	–0.049844	–0.073100

Source: Authors.

Notes: *** and * indicate significance at 1% and 10% levels, respectively.

Table 10.

Volatility Spillovers between Stock and GBPR during the Crisis Period

Coefficients	GARCH (1,1)		EGARCH (1,1)
Coefficients	Stock → GBPR	GBPR → Stock	Stock → GBPR	GBPR → Stock
α ₀	–9.76	0.000992	–0.00048	0.001281
α ₁	0.01687	–0.02668	0.020131	–0.05787
β ₀	1.08	4.70*	–0.05145	–0.6252**
β ₁	0.058191***	0.108273**	–0.01947	0.042814***
β ₂	0.931852***	0.826827***	0.993314***	0.941357***
γ			–0.06778***	–0.071840
θ	0.000146	–0.0063	0.244757	–12.2581***
ARCH LM	0.005296	–0.039694	0.039445	–0.054367

Source: Authors.

Notes: *** and ** indicate significance at 1% and 5%, respectively.

Table 11.

Volatility Spillover between Stock and GBPR during the Post-crisis Period

Coefficients	GARCH (1,1)		EGARCH (1,1)
Coefficients	Stock → GBPR	GBPR → Stock	Stock → GBPR	GBPR → Stock
α ₀	0.000146	0.000608**	0.000121	0.000413
α ₁	–0.12793***	–0.46945***	–0.13245***	–0.45381***
β ₀	1.28***	1.75**	–0.27263***	–0.27594***
β ₁	0.04826***	0.04396***	0.089602***	0.076922***
β ₂	0.912566***	0.939928***	0.980374***	0.976571***
γ			–5.78	–0.07398***
θ	–0.0001***	0.000228	–4.35348***	2.049775
ARCH LM	0.018374	–0.044345	0.023282	–0.052619

Source: Authors.

Notes: ***, ** and * indicate significance at 1%, 5% and 10% levels, respectively.

Volatility Spillover between Stock and YENR

During the pre-crisis period (see Table 12), both ARCH and GARCH terms are significant, it indicates volatility clustering and persistence. The θ is significant only in case of spillover from stock to YENR in both equations. But during the crisis period (see Table 13), there is no spillover in GARCH (1,1) model from stock to YENR. But there is unidirectional spillover from YENR to stock in GARCH model, and in EGARCH model there is bi-directional spillover. While in the post-crisis period (see Table 14), the θ is highly significant in both GARCH and E-GARCH, so there is bi-directional spillover from YENR to stock. And there is leverage effect in both, but the co-efficient is positive. So the good news has more impact as compared to bad news.

Table 12.

Volatility Spillovers between Stock and YENR during the Pre-crisis Period

Coefficients	GARCH (1,1)		EGARCH (1,1)
Coefficients	Stock → YENR	YENR → Stock	Stock → YENR	YENR → Stock
α ₀	0.000565	0.001584**	0.000722**	0.001198
α ₁	–0.25728***	–1.08856***	–0.23006***	–1.08345***
β ₀	1.92*	9.22*	–0.63034***	–1.0988***
β ₁	0.048627	0.133955***	–0.05297***	0.178429***
β ₂	0.630189***	0.838196***	0.931525***	0.895439***
γ			0.149746***	–0.15606***
θ	–0.00046***	0.000776	–3.40335***	6.208938
ARCH LM	–0.023547	0.071091	–0.020394	–0.028326

Source: Authors.

Notes: ***, ** and * indicate significance at 1%, 5% and 10% levels, respectively.

Table 13.

Volatility Spillovers between Stock and YENR during the Crisis Period

Coefficients	GARCH (1,1)		EGARCH (1,1)
Coefficients	Stock → YENR	YENR → Stock	Stock → YENR	YENR → Stock
α ₀	0.000822	0.002371**	0.000381	0.001099
α ₁	–0.27773***	–1.15092***	–0.28951***	–1.18845***
β ₀	3.03*	–1.91	–0.08046***	–0.00366
β ₁	0.073288**	–0.0059	–0.02892***	0.00109
β ₂	0.892264***	0.951893***	0.989581***	0.964753***
γ			–0.10884***	–0.05727**
θ	–0.00022	–0.00157**	0.77285*	–4.1167***
ARCH LM	0.020492	–0.037197	0.111360	–0.036036

Source: Authors.

Notes: ***, ** and * indicate significance at 1%, 5% and 10% levels, respectively.

Table 14.

Volatility Spillovers between Stock and YENR during the Post-crisis Period

Coefficients	GARCH (1,1)		EGARCH (1,1)
Coefficients	Stock → YENR	YENR → Stock	Stock → YENR	YENR → Stock
α ₀	–7.57	0.000399	8.49	0.000388
α ₁	–0.29972***	–0.57316***	–0.30318***	–0.61945***
β ₀	4.26E–06***	1.32***	–0.78053***	–13.8125***
β ₁	0.087346***	0.029699***	0.081439***	0.06397
β ₂	0.83467***	0.955409***	0.935074***	–0.47866**
γ			0.063568***	0.006974
θ	–0.00021***	0.000258***	–3.81493***	–11.3444***
ARCH LM	0.046346	–0.048423	0.054036	0.006649

Source: Authors.

Notes: ***, ** and * indicate significance at 1%, 5% and 10% levels, respectively.

Conclusion

Our study examines the volatility spillover between Indian equity and foreign exchange market in during, pre and post subprime crisis periods by taking daily returns of Nifty and three exchange rates, namely USDR, GBPR and YENR. Initially, the volatility clustering is observed, and in overall, both stock and foreign exchange markets exhibit volatility persistence and clustering (past studies show similar results, see Bal, 2016; Bal & Singhraul, 2016). We provide mix evidence of asymmetric volatility spillover between these markets and it is time varying in nature. There is bi-directional spillover between the stock market and USD in both GARCH (1,1) and EGARCH (1,1) in the pre-crisis period, while unidirectional in other cases. We found surprise result in case of stock and GBP, that is, no spillover between these markets in pre-crisis period. When it comes to during the crisis period, the results of EGARCH model show bi-directional volatility transmission between stock and exchange rates except GBP (for GBP only unidirectional spillover in EGARCH). Immediately after crisis period, it has been observed that there is bi-directional spillover between stock and YENR, and unidirectional spillover in other case, that is, from USD to stock and stock to GBP. Except during the crisis period, in both pre and post crisis, there is leverage effect in case of USD and stock. In both pre- and post-crisis periods, from stock to USD, the coefficient is positive while it is negative from USD to stock. In case of GBP, in both during and post-crisis periods, there is negative coefficient in case of both way transmissions. In case of yen, same effect has been observed in both pre- and post-periods, that is, negative from yen to stock and positive from stock to yen. We also support the findings of Panda and Deo (2014) that there is a strong relationship between the Indian stock prices and USD exchange rates as compared to others. Our findings also support the view that the Indian market did not get affected by the financial crisis to a great extent. Moreover, we observed mix evidence of spillover in the post-crisis period. This may be because of the introduction of new derivative instruments after the crisis period, and implementation of new policy measures for stabilizing the market. The past studies also experienced similar observation (see Ghosh, 2014). Thus, quick recovery from the crisis impact indicates that the Indian market is no more juvenile and there is significant improvement in the market micro structure.

Managerial Implication

The contribution of our study has several dimensions: first of all it will contribute to the existing literature in this area. And in the present business environment, the study will help many hedgers, arbitrators and companies in diversifying their forex risk and it will also help in portfolio diversification of the companies in international market. The findings of the study will also help the policy makers in this regard.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

Footnotes

Acknowledgements

The authors are grateful to the anonymous referees and editor of the journal for their extremely useful suggestions to improve the quality of the article. Usual disclaimers apply.

References

Apte

(2001). The interrelationship between stock markets and the foreign exchange market (Working Paper No. 169). Bangalore: IIM Bangalore. Retrieved from http://www.iimb.ernet.in/research/sites/default/files/wp.iimb_.169.pdf

Badrinath

H. R.

& Apte

P. G.

(2005). Volatility spillovers across stock, call money and foreign exchange markets. Retrieved from http://www.nse-india.com/content/research/comppaper109

Bal

G. R.

(2016). Revisiting random walk behavior of stock market: Evidence from Indian market. GE-International Journal of Management Research, 4(8), 80–87.

Bal

G. R.

& Singhraul

B. P.

(2016). Efficiency and volatility of foreign exchange market: Evidence from selected exchange rates. Pacific Business Review International, 1(4), 96–101.

Bollerslev

(1986). Generalised autoregressive conditional heteroskedasticity. Journal of Econometrics, 31(3), 307–327.

Bonga-Bonga

& Hoveni

(2013). Volatility spillovers between the equity market and foreign exchange market in South Africa in the 1995–2010 Period. South African Journal of Economics, 81(2), 260–274. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1111/saje.12001

Chittedi

(2015). Financial crisis and contagion effects to Indian Stock Market: ‘DCC-GARCH’ analysis. Global Business Review, 16(1), 50–60. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1177/0972150914553507

Engel

R. F.

(1982). Auto regressive conditional heteroskedasticity and estimates of the variance of UK inflation. Econometrica, 50(4), 987–1008.

Francis

B. B.

Hunter

D. M.

& Hasan

(2002). Return volatility linkages in international equity and currency markets (Bank of Finland Discussion Papers No. 9-2002). Helsinki: Bank of Finland.

10.

Ghosh

(2014). Volatility spillover in the foreign exchange market: The Indian experience. Macroeconomics and Finance in Emerging Market Economies, 7(1), 175–194. doi: 10.1080/17520843.2013.856334

11.

Gupta

& Kamila

(2015). Dynamic linkages between implied volatility indices of developed and emerging financial markets: An econometric approach. Global Business Review, 16(5S), 46S–57S. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1177/0972150915601237

12.

Kanas

(2000). Volatility spillovers between stock returns and exchange rate changes: International evidence. Journal of Business Finance & Accounting, 27(3–4), 447–467. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1111/1468-5957.00320

13.

Kumar

(2014). Correlations, return and volatility spillovers in Indian exchange rates. Global Business Review, 15(1), 77–91. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1177/0972150913515577

14.

Md-Yusuf

& Rahman

(2013). Granger causality relationship between Malaysia equity market and exchange rate volatility. International Journal of Trade, Economics and Finance, 19–24. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.7763/ijtef.2013.v4.254

15.

Nelson

D. B.

(1991). Conditional heteroscedasticity in asset returns: A new approach. Econometrica, 59(2), 347–370.

16.

Panda

& Deo

(2014). Asymmetric cross-market volatility spillovers: Evidence from Indian equity and foreign exchange markets. Decision, 41(3), 261–270. doi: 10.1007/s40622-014-0044-2.pdf

17.

Saha

& Chakrabarti

(2011). Financial crisis and financial market volatility spill-over. International Journal of Applied Economics and Finance, 5(3), 185–199.

18.

Sahoo

(2012). Volatility transmission in the exchange rate of the Indian rupee (RBI Working Paper Series P S (DEPR): 08/2012). India: RBI.

19.

Srivastava

Bhatia

& Gupta

(2015). Financial crisis and stock market integration: An analysis of select economies. Global Business Review, 16(6), 1127–1142. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1177/0972150915604519

20.

Zabiulla (2015). Volatility clustering and leverage effect in the Indian Forex Market. Global Business Review, 16(5), 785–799. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1177/0972150915591453